Thoughts from me on your outstanding items. Copying the mailing list to get more input.
> Integrate into Fennec One of the goals you mentioned this morning is to continue being able to ship from your Git repo. I'd suggest one of these approaches: * Source-based: deliver the source for the library part into mozilla-central from a script (just as we do for android-sync), and integrate it into mobile/android's own build system bits. This becomes an integral part of Fennec, just like Sync, but you develop it upstream. This doesn't provide you with an easy way to localize MozStumbler (standalone) itself using Mozilla's existing l10n patterns; I don't know if you've already solved l10n, don't care at this stage, or something else. * All source-based: dump all the source into m-c. Just like the former, but you do the additional work to build the standalone app out of m-c. This is probably a lot more work, but more payoff, too. * Artifact-based: dump new jars into m-c. We'd only take this approach if the source-based approach was somehow impossible, or if we needed a temporary shortcut. nalexander is the authority here, and he's very happy to help! > no new permissions? Ideally. If new permissions are needed, we ordinarily coordinate with other features that need permissions, which will affect schedule. There's a high bar for adding permissions, because they dramatically affect user upgrades. > we ask 1 question of the user in Android Notification screen, can we ask a > 2nd question? Paging ibarlow/alam/yuan. We already show a Data Reporting notification, and probably location stumbling shouldn't be on by default. Adding a second notification is likely a step too far, so I'd suggest a call to action from a snippet, or a non-first-run nudge. Provide both incentive and opportunity. > Sync code needs a peer-review Feel free to flag me for r?. This is definitely my ballpark. Where's your bug component? If you don't have one yet, let's fix that. We might also consider breaking some parts of Android Background Services out of android-sync (and some parts of Fennec, too!). We're past the point of needing some kind of shared library, I think. (Nick and I should chat about that.) > run stumbler service as a separate process, Android *should* be less likely > to kill it. Maybe. It might be more likely to kill it: better it kills the stumbler than Fennec. And remember that having a third process is expensive, and makes it more likely that anything will be killed at all! Best practice is to simply not use extra processes unless you have some clear reason to do so (e.g., loading NSS for the passwords provider). _______________________________________________ mobile-firefox-dev mailing list [email protected] https://mail.mozilla.org/listinfo/mobile-firefox-dev
